Investigation of first-order corrections in bistatic radiative transfer models for remote sensing of vegetated terrain

Abstract

The work is carried out in the field of satellite based remote sensing of soil moisture on a global-scale. With ongoing progress in the improvement of the resolution of satellite based scatterometers in the microwave domain, current techniques for the retrieval of soil-moisture content from active microwave observations are to be revised in order to include in greater detail the effects induced by different roughness conditions of the soil surface and the attenuation of the radiation resulting from the propagation through the vegetation coverage. The propagation of an incoherent beam of radiation from the transmitter to the surface (penetrating the vegetation-cover) and back to the detector is described by means of the Radiative Transfer Theory. In order to be able to derive an analytic expression for the backscattered radiation, it is assumed that the vegetation coverage can be treated as a weakly scattering tenuous distribution of particles for impinging radiation at frequencies in the microwave domain. Applying this assumption, the solution to the radiative transfer equation (describing the backscattered radiation) can be expressed in terms of a successive orders of scattering series, where the individual terms represent radiation that has been scattered n-times either by the ground surface or by the vegetation-coverage. Current models usually incorporate the effects induced by a vegetation-coverage via the so-called Water Cloud-Model which is based on a zero-order approximation of the successive orders of scattering series. The "Cloud Model" thus assumes that multiple scattering terms are negligible. The validity of this assumption however is highly dependent on the given estimate of the "optical depth" and the "single scattering albedo" ! of the considered vegetation as well as the directionality of the scattering-distributions and might not be reliable in general. While some authors incorporate empirically driven corrections to the cloud-model approximation in order to account for multiple-scattering effects, a general physically based treatment of the first-order scattering contributions is to our knowledge currently only available via numerical simulations. The proposed thesis is therefore intended to find methods that allow an estimation of the interaction-contributions in an analytic manner taking into account both the directionality of the surface reflectance as well as the directionality of the scattering phase function of the vegetation-constituents. The behaviour of the model is intended to be tested against angular dependent backscatter measurements provided by the Advanced Scatterometer (ASCAT) instrument as flown on the Metop-A satellite launched in 2006.